Large Scale CO Observations of a Far-Infrared Loop in Pegasus; Detection of a Large Number of Very Small Molecular Clouds Possibly Formed via Shocks

0.45 MB·English

by  Hiroaki Yamamoto

Large Scale CO Observations of a Far-Infrared Loop in Pegasus;
Detection of a Large Number of Very Small Molecular Clouds
Possibly Formed via Shocks
H. Yamamoto1, A. Kawamura1, K. Tachihara2, N. Mizuno1, T. Onishi1 and Y. Fukui1
6
0 ABSTRACT
0
2

n WehavecarriedoutlargescaleCOobservationswithamm/sub-mmtelescope
a
J NANTEN toward a far infrared loop-like structure whose angular extent is about

5 20×20 degrees around (l, b) ∼ (109◦, −45◦) in Pegasus. Its diameter corresponds
1
to ∼ 25 pc at a distance of 100 pc, adopted from that of a star HD886 (B2IV)

1 near the center of the loop. We covered the loop-like structure in the 12CO (J =
v
5 1–0) emission at 4′–8′ grid spacing and in the 13CO (J = 1–0) emission at 2′ grid
1
spacing for the 12CO emitting regions. The 12CO distribution is found to consist
3
1 of78smallclumpycloudswhosemassesrangefrom0.04M to11M ,and∼83%
⊙ ⊙
0
6 of the 12CO clouds have very small masses less than 1.0 M⊙. 13CO observations
0 revealed that 18 of the 78 12CO clouds show significant 13CO emission. 13CO
/
h
emission was detected in the region where the column density of H derived from
p 2
- 12CO is greater than 5×1020 cm−2, corresponding to Av of ∼ 1 mag, which takes
o
r into account that of HI. We find no indication of star formation in these clouds
t
s in IRAS Point Source Catalog and 2MASS Point Source Catalog. The very low
a
: mass clouds, M ≤ 1M , identified are unusual in the sense that they have very
v ⊙
i weak 12CO peak temperature of 0.5 K–2.7 K and that they aggregate in a region
X
r of a few pc with no main massive clouds; contrarily to this, similar low mass
a
cloudslessthan1M inotherregionspreviously observed including thoseathigh
⊙
Galactic latitude are all associated with more massive main clouds of ∼ 100 M .
⊙
A comparison with a theoretical work on molecular cloud formation (Koyama &
Inutsuka 2002) suggests that the very low-mass clouds may have been formed in
the shocked layer through the thermal instability. The star HD886 (B2IV) may
be the source of the mechanical luminosity via stellar winds to create shocks,
forming the loop-like structure where the very low-mass clouds are embedded.
1
Department of Astrophysics, Nagoya University, Chikusa-ku, Nagoya, Japan 464-8602;
[email protected]
2Graduate School of Science and Technology, Kobe University, 1-1 Rokko-dai, Nada-ku, Kobe, Japan
657-8501
– 2 –
Subject headings: ISM: clouds — ISM: individual(High Latitude Clouds) —radio
lines: ISM — stars: formation — stars: winds
1. INTRODUCTION
High Galactic latitude molecular clouds (hereafter HLCs) are typically located at |b|
& 20◦–30◦. Since the Gaussiun scale height of CO is estimated to be ∼ 100 pc in the
inner Galactic disk (e.g., Magnani et al. 2000), HLCs are likely located very close to the
Sun, within a few hundred pc or less. Their proximity to the Sun and the low possibility
of overlapping with other objects along the line of sight enable us to study them with a
high spatial resolution and to compare CO data unambiguously with the data at other
wavelengths. HLCs have lower molecular densities compared with dark clouds where the
optical obscuration is significant. Therefore, HLCs are often called as translucent clouds
(e.g., van Dishoeck & Black 1988) and most of the known HLCs are not the sites of active
star formation, although a few of them are known to be associated with T Tauri stars (e.g.,
Magnani et al. 1995; Pound 1996; Hearty et al. 1999).
Given the very small distances of HLCs, it is a challenging task for observers to make
a complete survey for HLCs over a significant portion of the whole sky. 12CO (J = 1–
0) emission has been used to search for HLCs because the line emission in the mm band is
strongestamongthethermallyorsub-thermallyexcitedspectrallinesofinterstellarmolecular
species. It is however difficult to cover an area as large as tens of square degrees subtended
by some of the HLCs because of the general weakness of the 12CO emission, typically ∼ a
few K (e.g., Magnani et al. 1996), with existing mm-wave telescopes in a reasonable time
scale. HLCs have been therefore searched for by employing various large-scale datasets at
other wavelengths including the optical obscuration (Magnani et al. 1985; Keto & Myers
1986), the infrared radiation (Reach et al. 1994), and the far-infrared excess over HI (=FIR
excess)(Blitz et al. 1990; Onishi et al. 2001). On the other hand, unbiased surveys in CO at
high Galactic latitudes have been performed at very coarse grid separations of 1◦ resulting
in a small sampling factor of a few % (Hartmann et al. 1998; Magnani et al. 2000). Most
recently, Onishi et al. (2001) discovered 32 HLCs or HLC complexes. This search was made
based on the FIR excess, demonstrating the correlation among FIR excess clouds with CO
clouds is a useful indicator of CO HLCs.
Previous CO observations of individual HLCs at higher angular resolutions show that
HLCs exhibit often loop-like or shell-like distributions having filamentary features with
widths of several arc min or less (Hartmann et al. 1998; Magnani et al. 2000; Bhatt 2000),
and in addition that HLCs often compose a group, whose angular extent is ∼ 10 degrees or
– 3 –
larger. In order to better understand the structure of HLCs and to pursue the evolution of
HLC complexes, CO observations covering tens of square degrees at a high angular resolu-
tion are therefore crucial. The past observations of such complexes of HLCs are limited to
a few regions including Polaris flare (Heithausen & Thaddeus 1990), Ursa Major (Pound &
Goodman1997) and the HLC complex toward MBM 53, 54, and 55 (Yamamoto et al. 2003).
Pound & Goodman (1997) showed an arc-like structure of the molecular cloud system and
suggested that the origin of such structures could be some explosive events. Most recently,
Yamamoto et al. (2003) carried out extensive observations of the molecular cloud complex
including MBM 53, 54, and 55 and suggest that the HLCs may be significantly affected by
past explosive events based on the arc-like morphologies of molecular hydrogen (see also Gir
et al. 1994).
The region of MBM 53, 54, and 55 is of particular interest among the three, because it
is associated with a large HI cloud of ∼ 590 M⊙ at a latitude of −35 degrees and because
there is a newly discovered HLC of 330M⊙, HLCG92−35, which is significantly HI rich with
a mass ratio M(H )/M(HI) of ∼ 1, among the known HLCs (Yamamoto et al. 2003). This
2
cloud was in fact missed in the previous surveys based on optical extinction (Magnani et al.
1985). Subsequent to these observations we became aware of that the region is also very rich
in interstellar matter as shown by the 100µm dust features (Kiss et al. 2004). There is a
loop-like structure shown at 100 µm around (l, b) ∼ (109◦, −45◦). Toward the center of the
loop, an early type star HD886(B2IV) is located and may play a role in creating the loop.
Its proper motion is large at a velocity of a few km s−1, suggesting that the stellar winds of
the star might have continued to interact with the surrounding neutral matter over a few
tens of pc in ∼ a few Myr. Magnani et al. (1985) and Onishi et al. (2001) yet observed
only a small part of this region. In order to reveal the large scale CO distribution of the
region, we have carried out observations toward (l, b) ∼ (109◦, −45◦) by 12CO (J = 1–0) and
13CO (J = 1–0) with NANTEN 4-meter millimeter/sub-mm telescope of Nagoya University
at Las Campanas, Chile. We shall adopt the distance of 100 pc from the sun to the loop-like
structure which is equal to the distance of the B2 star in the center of the loop, and is also
a typical value for the HLCs.
2. OBSERVATIONS
12CO (J = 1–0)and 13CO (J = 1–0)observations were made with the 4-meter telescope,
NANTEN, of Nagoya University at Las Campanas Observatory of Carnegie Institutions of
Washington, Chile. The front-end was an SIS receiver cooled down to 4 K with a closed-
cycle helium gas refrigerator (Ogawa et al. 1990). The backend was an acousto-optical
– 4 –
spectrometer with 2048 channels, and the total bandwidth was 40 MHz. The frequency
resolution was 35 kHz, corresponding to a velocity resolution of ∼ 0.1 km s−1. A typical
system noise temperature was ∼ 200 K (SSB) at 115.271 GHz and ∼ 150 K (SSB) at 110.201
GHz. The half-power beam width was about 2.′6, corresponding to 0.076 pc at a distance of
100 pc. The pointing accuracy was better than 20′′, as established by radio observations of
Jupiter, Venus, and the edge of the Sun in addition to optical observations of stars with a
CCD camera attached to the telescope.
Theobservedregionin12COwas∼240squaredegreestowardthewholeareaoftheloop-
like structure centered at around (l, b) ∼ (109◦, −45◦) shown in a 100 µm map by Schlegel
et al. (1998). First, the 12CO observations were made at a grid spacing of 8′×cos(b) and 8′
in Galactic longitude and latitude, respectively. Then, the regions where the 12CO emission
is significantly detected were observed at a grid spacing of 4′×cos(b) and 4′ in Galactic
longitude and latitude, respectively. The 13CO observations were made in and around the
whole area where the peak temperature of 12CO emission is higher than 2.0 K at a grid
spacing of 2′×cos(b) and 2′ in Galactic longitude and latitude, respectively. The periods of
12CO observations were several sessions between 2002 May and November and those of 13CO
were those between 2003 April and August. All the observations were made by frequency
switching whose interval is 20 MHz, corresponding to ∼ 50 km s−1. The integration times
per point of 12CO and 13CO observations were typically ∼ 30 s and ∼ 75 s, respectively,
resulting in typical rms noise temperatures per channel of ∼ 0.35 K and ∼ 0.15 K in the
radiation temperature, T∗, respectively. In reducing the spectral data, we subtracted forth-
R
order polynomials for the emission-free parts in order to ensure a flat spectral baseline. Total
numbers of observed points of 12CO and 13CO are 16890 and 3100, respectively.
We employed a room-temperature blackbody radiator and the sky emission for the
intensity calibration. An absolute intensity calibration and the overall check of the whole
system were made by observing Orion KL [α(1950) = 5h32m47.s0, δ(1950) = −5◦24′21′′]
every 2 hours. We assumed the T∗ of Orion KL to be 65 K for 12CO and 10 K for 13CO.
R
3. RESULTS
3.1. 12CO Observation
3.1.1. Distribution and Past Detection of 12CO Clouds
Figure 1 shows the distribution of the velocity-integrated intensity map of 12CO emis-
sion. Wedefined a 12CO cloud as a collection of morethan two contiguous observed positions
– 5 –
whose integrated intensity exceeds 0.77 K km s−1 (5σ). Based on the definition, we iden-
tified 78 molecular clouds in this region. Molecular clouds are concentrated from (l, b) ∼
(107◦, −37◦) to (116◦, −45◦) and around (l, b) ∼ (114◦, −52◦). Most of the molecular
clouds are very small, having size of . 1◦. Figure 2 shows the distribution of the CO su-
perposed on the SFD 100 µm (Schlegel et al. 1998), which was derived from a composite of
the COBE/DIRBE and IRAS/ISSA maps, with the foreground zodiacal light and confirmed
point sources removed. CO clouds are distributed along the infrared loop whose diameter is
∼ 25 pc. We detected little CO emission within the loop-like structure, while toward some
of the local peaks of SFD 100 µm there is no CO emission.
Figure 3 shows the peak radial velocity distribution derived from the present 12CO data
set. The velocity in Figure 3 is derived by a single gaussiun fitting from all CO spectra. The
velocity range of the molecular clouds is from −18.3 km s−1 to 0.3 km s−1 and there is no
systematic large scale velocity gradients.
Someof themolecular cloudshave alreadybeenknown byprevious observations. Molec-
ular clouds toward (l, b) ∼ (110◦.18, −41◦.23) and (117◦.36, −52◦.28) are identified by Magnani
et al. (1985) and named as MBM 1 and MBM 2, respectively. DIR117−44 and DIR105−38
identified by Reach et al. (1998) are also identified in CO toward (l, b) ∼ (116◦.5, −44◦.0)
and (105◦.0, −38◦.0) by Onishi et al. (2001). Magnani et al. (1986) detected CO emission at
(l, b) ∼ (112◦, −40◦). Magnani et al. (2000) also covered this region even though they made
observations on a locally Cartesian grid with 1◦(true angle) spacing in longitude and lati-
tude for a beam size of 8.′8, they detected CO emission at eight positions of (l, b) ∼ (103◦.2,
−38◦.0), (103◦.2, −39◦.0), (104◦.4, −39◦.0), (106◦.8, −37◦.0), (108◦.0, −52◦.0), (109◦.5, −51◦.0),
(110◦.4, −41◦.0), and (111◦.0, −50◦.0) in the present region while they missed the present small
molecular clouds whose sizes are less than several arc min in Figure 1 due to the coarse grid
spacing.
3.1.2. Physical Properties of 12CO Molecular Clouds
Seventy-eight 12CO molecular clouds are identified in the present region. For each
molecular cloud, ∆V derived from single Gaussian fitting was from 0.5 to 3.7 km s−1, and
the radial velocity, V , ranges from −15.7 to −0.1 km s−1. The maximum brightness
LSR
temperature, T∗(12CO) ranges from 0.5 to 5.7 K. The radius of a cloud, R, which is defined
R
as the radius of an equivalent circle having the same area, i.e., and R(pc)= A/π where A
is the total cloud surface area within the 5σ–contour level, ranges from 0.07 to 0.79 pc. The
p
peak column density of molecular hydrogen, N(H ), in each cloud derived by assuming a
2
conversion factor of 1.0×1020 cm−2/(K km s−1) (Magnani et al. 2000) ranges from 8.0×1019
– 6 –
to 1.7×1021 cm−2 with the present detection limit, 7.7×1019 cm−2, corresponding to mass
detection limit of 0.014 M . We estimate the molecular mass, M(12CO), by using the
⊙
following formula
M(12CO) = µm Σ[D2ΩN(H )], (1)
H 2
where µ is the mean molecular weight, assumed to be 2.8 by taking into account a relative
helium abundance of 25% in mass, m is the mass of the atomic hydrogen, D is the distance
H
from the Sun to the molecular clouds, and Ω is the solid angle subtended by a unit grid
spacing of (4′)×(4′×cos(b)). M(12CO) ranges from ∼ 0.04 to ∼ 11 M and the total mass
⊙
of molecular clouds is ∼ 64 M . These physical properties are listed in Table 1 and the
⊙
histograms of T∗(12CO), ∆V, log(R), and log(N(H )) of these clouds are shown in Figure
R 2
4. Histograms in Figure 4 are divided into three different categories, Usual Cloud (hereafter
UC) whose mass is greater than 1 M , Small Cloud (hereafter SC) whose mass is between
⊙
0.1 and 1 M , and Very Small Cloud (hereafter VSC) whose mass is less than 0.1 M . It
⊙ ⊙
is remarkable that there are a number of molecular clouds having mass less than 1 M and
⊙
that the fractions of SC and VSC are 43/78 ∼ 55% and 22/78 ∼ 28% in the present region,
respectively. In addition, the sizes of SC and VSC are equal to or less than 0.1 pc. We also
note that the peak temperatures of SC and VSC are typically in a range from 0.5 K to 2.7
K, well below that of UC in the same region.
3.2. The Detection and Physical Properties of the 13CO Molecular Clouds
Figure 5 shows the distribution of the velocity-integrated intensity map of the 13CO
emission superposed on the 12CO distribution. The total area of the 13CO observations is ∼
29 square degrees toward 38 of the 78 12CO clouds. We observed all of 13 UCs, 24 of 43 SCs,
and 3 of 22 VSCs. We detected 13CO emission at 11 of the 13 UCs, 8 of the 24 SCs, and
none of the 3 VSCs, indicating a trend that the 13CO intensity increases with 12CO cloud
mass.
A 13CO cloud is defined in the same way as for a 12CO cloud except for the lowest
integrated intensity level, 0.3 K km s−1 (3σ). Based on the definition, we identified 33 13CO
clouds. For the 33 13CO molecular clouds, ∆V derived from single Gaussian fitting is ∼ 1.5
km s−1 and V of them ranges from −13.1 to −1.9 km s−1. Other physical properties, the
LSR
maximum brightness temperature, T∗(13CO),andRrangefrom0.3to2.3K andfrom0.04to
R
0.21 pc, respectively. The physical parameters including the molecular column density and
mass (hereafter M ) are derived on the assumption of local thermodynamic equilibrium
LTE
(LTE). To derive the column density of molecular hydrogen, the optical depth of 13CO is
– 7 –
estimated by using the following equations,
T∗(13CO) 1 −1
τ(13CO) = ln 1− R −0.164 , (2)
5.29 exp(5.29/T )−1
" (cid:26) ex (cid:27) #
where T is the excitation temperature of the J = 1–0 transition of CO in K andwas derived
ex
from
5.53
T = . (3)
ex ln{1+5.53/[T∗(12CO)+0.819]}
R
T was estimated to be 9.4 K from our 12CO data. The 13CO column density, N(13CO), is
ex
estimated by
τ(13CO)T (K)∆V(kms−1)
N(13CO) = 2.42×1014 × ex (cm−2). (4)
1−exp[−5.29/T (K)]
ex
The ratio of N(H )/N(13CO) was assumed to be 7×105 (Dickman 1978). The M of a
2 LTE
cloud from N(H ) is derived by the same way as 12CO (see equation (1)). The column
2
density and M range from 2.3×1020 to 1.7×1021 cm−2 and 0.03 to 1.41 M , respectively,
LTE ⊙
where the detection limit in the column density is 2.0×1020 cm−2, coressponding to mass
limit of 0.009 M , smaller than that of 12CO because the observations of 13CO were made
⊙
by higher grid sampling and lower rms noise fluctuations than those of 12CO, respectively.
Figure 6 shows the histograms of each physical property. The virial mass, M , of a cloud
vir
was derived by using the following equation, assuming isothermal, spherical, and uniform
density distribution with no external magnetic pressure:
M = 209×R×∆V2 , (5)
vir comp
where R and ∆V are the radius (pc) and line width (km s−1) of the composite profile
comp
obtained by averaging all the spectra within a cloud, respectively (for details of the line
width of composite profiles, see Yonekura et al. 1997; Kawamura et al. 1998). From this
equation, M is estimated to be in a range from 4.7 to 197 M . These physical properties
vir ⊙
are also listed in Table 2.
4. CORRELATIONS AMONG THE CLOUD PHYSICAL PARAMETERS
4.1. Mass Spectrum and Size Linewdith Relation
Figure 7a and 7b show the mass spectrum of the present 12CO and 13CO clouds. The
spectra have been fitted by the maximum-likelihood method (Crawford et al. 1970), and it
is found that they are well fitted by a single power law as follows; dN/dM ∝ M−1.53±0.13 for
– 8 –
the 12CO clouds and dN/dM ∝ M−1.36±0.10 for the 13CO clouds. These values of the spectral
indices seem to be similar to those for the higher mass range (e.g., Yonekura et al. 1997).
Figure 8 shows a plot of size, R, versus line width, ∆V, of the 13CO clouds in this
region and for a comparison with other HLCs, MBM 53, 54, and 55 complex (Yamamoto
et al. 2003). We can make fitting as follows by using a least-squares fitting, log(∆V)
= (0.22±0.43) × log(R) + (0.37±0.52) (c.c.=0.23) for the present region and log(∆V) =
(0.43±0.32) × log(R) + (0.53±0.28) (c.c.=0.37) for MBM 53, 54, and 55 complex. The low
correlation coefficient (c.c.) indicates that there is no significant correlation between R and
∆V because of a small range of R. Here we do not show the same relationship for 12CO,
because the non-circular shape of the 12CO clouds may not be appropriate to derive reliable
R.
4.2. M vs. M
LTE vir
Figure9showsaplotofM versusM . Thepresent13COcloudsarelocatedfarabove
LTE vir
the equilibrium line where M is equal to M , indicating that the 13CO clouds are not in
LTE vir
the virial equilibrium. This indicates that none of the molecular clouds are gravitationally
bound. These parameters can be fitted by using a least-squares fitting as follows, log(M )
vir
= (0.91±0.30) × log(M ) + (2.23±0.29) (c.c.=0.66) for present molecular clouds and
LTE
log(M ) = (0.77±0.13) × log(M ) + (0.16±0.08) (c.c.=0.74) for MBM53, 54, and 55
vir LTE
complex. As mentioned in Yamamoto et al. (2003), the present molecular clouds also
tend to be more virialized as the mass increases. For the Gemini and Auriga, and Cepheus-
Cassiopeia regions, theindices ofM forM of13COclouds areestimated tobe0.72±0.03
LTE vir
and 0.62±0.03 for the cloud mass range of M < 104 M and 102 M < M < 105 M ,
LTE ⊙ ⊙ LTE ⊙
respectively (Kawamura et al. 1998; Yonekura et al. 1997). Although the mass ranges of
MBM 53, 54, and 55 complex and of this region are 10−1 M < M < 102 M and 10−2
⊙ LTE ⊙
M < M < 1 M , respectively, difference in the power-law indices among these regions
⊙ LTE ⊙
is small and a tendency that the SCs have large ratios of M /M is commonly seen.
vir LTE
5. COMPARISON WITH OTHER WAVELENGTH DATA
5.1. No Sign of Star Formation
In order to look for sings of star formation associated with the present molecular clouds,
we searched the IRAS point source catalog for candidates of protostellar objects satisfying
the following criteria: (1) point sources having a data quality flag better than 2 in 4 bands,
– 9 –
(2) flux ratios at 12, 25, 60 µm satisfying both log(F /F ) < −0.3 and log(F /F ) < 0,
12 25 25 60
and not identified as galaxies or planetary nebulae and stars. We find that there are no
cold IRAS point sources satisfying these criteria in the present molecular clouds. We also
find that there are no IRAS point sources having a spectrum like a T-Tauri type star or no
YSOs identified from Point Source Catalog of Two-Micron All-Sky Survey in this region.
Here we select the 2MASS sources whose signal to noise ratio of valid measurements in all
bands are greater than 10 and extract the sources which have the spectra like T Tauri stars
in (J−H)–(H−K) color–color diagram (e.g., Meyer et al. 1997) . These results suggest that
the present molecular clouds are not the site of recent star-formation, or that the region is
not remnants of past star formation. Such a low level of star formation is similar to the
other HLCs including MBM 53, 54, and 55 complexes.
5.2. Comparison with HI
Figure 10 shows the integrated intensity map of HI taken from a Leiden-Dwingeloo
HI survey (Hartmann & Burton 1997) superposed on the integrated intensity of CO. The
integrated velocity range is from −16 to 0 km s−1, corresponding to the velocity range of the
12CO emission. Because an angular resolution of ∼ 30′ is coarser than that of the present
CO observations by a factor of ∼ 10, we discuss here only the overall comparison between
CO and HI distributions. The HI distribution is loop-like and the molecular clouds are
distributed nearly along the HI loop.
Figure 11 shows the position-velocity diagram of HI integrated from −40◦.5 to −39◦.5
and −52◦.5 to −51◦.5 in Galactic latitude, respectively. The hole like structures can be seen in
HI, suggesting that these HI clouds are expanding. These expanding structures in Figures
11(a) and 11(b) correspond to the Galactic northern and southern HI shells shown as the
thick dashed (semi) ellipses in Figure 10, respectively, while the expanding motion is not
seen in CO (See Figure 3). These two expanding shells are also identified by an infrared
radiation (Kiss et al. 2004). From Figure 10, an HI cloud around (l, b) ∼ (109◦, −52◦)
seems to be located on the left side of the southern expanding shell and the molecular clouds
are distributed nearby. It is difficult to distinguish with which HI shell the molecular clouds
located around (l, b) ∼ (109◦, −52◦) are associated because the HI velocities of the two
shells are similar with each other. The shape of 12CO and SFD100 µm radiation around
(l, b) ∼ (109◦, −52◦) in Figure 2 is also similar to the left side of the expanding shell. If
this is true, there may be two expanding structures in the present region. HD886 (109◦.43,
−46◦.68) is located near the center of the northern expanding shell, indicating that HD886
may be affecting the northern expanding shell. The parallax of HD886 has been measured
– 10 –
to be 9.79±0.81 mas (Perryman et al. 1997), corresponding to a distance from the Sun of
102+9 pc. The proper motion has also been measured to be µα∗=0′.′0047 yr−1, µδ=−0′.′0082
−8
yr−1 by Perryman et al. (1997). From this proper motion, the velocity of HD886 in the L-B
map is estimated to be ∼ 4.3×10−6 pc yr−1 at ∼ 100 pc (see Figure 10). We use typical
values of the stellar wind for B2(IV) star on dM/dt=10−9 M yr−1 and V = 1000 km
⊙ ∞
s−1 (e.g., Snow 1982). From these parameters, the energy injected to the northern shell is
estimated to be ∼ 1047 ergs in a few ×106 yr. The expanding energy of the northern shell is
estimated to be ∼ 1048 ergs from the atomic and molecular hydrogen, using that the masses
of amomic and molecular hydrogen associated with the northern shell are ∼ 400 and 42 M ,
⊙
respectively, the expanding velocity is ∼ 7 km s−1 which is estimated from Figure 11 and
the equation of E =1/2MV2 . Since the expanding energy of the atomic and molecular
exp exp
hydrogen is comparable to the energy from HD886, additional source of energy other than
HD886 such as photo evaporation is needed to explain the expanding energy because the
energy conversion efficiency of the stellar wind is . 10%. The energy of the southern shell is
estimated to be ∼ 1047 ergs, by using that the masses of atomic and molecular hydrogen are
∼ 1000 and 16 M , respectively, and that the expanding velocity is ∼ 9 km s−1. We could
⊙
not find possible candidates of the energy source for the expanding feature in the literature
(SIMBAD) and there are no counterparts in optical or X-ray wavelength. Although we could
not identify the possible candidates, we cannot exclude a possibility that these objects may
have escaped from the region in a few × 106−7 yr after forming these structures.
6. DISCUSSION
6.1. Physical States of the Small Clouds
The present observations have revealed numerous molecular clouds having very small
mass of less than 1 M . It is of considerable interest to pursue the physical states of these
⊙
SCs from view points of cloud physics and chemistry as well as of the origin of molecular
clouds.
We shall hereafter focus on the low mass 12CO clouds whose mass is less than 1 M .
⊙
The total number of such clouds is 65 among 78. The 13CO emission has been searched
for toward 24 of the 43 12CO low-mass clouds whose mass is in a range of 0.1–1.0 M and
⊙
has been detected from 8 of them. Figure 12 shows correlations of the molecular column
density, estimated from 12CO and 13CO, of the clouds where both 12CO and 13CO emission
are detected. It is seen that almost all of the 12CO clouds having molecular column density
greater than 5×1020 cm−2, corresponding to the visual extinction of 0.55 mag if we use the
relationship of N(H )=9.4×1020×Av cm−2 (Bohlin et al. 1978; Hayakawa et al. 1999), show
2